Home About Login Current Archives Announcements Editorial Board
Submit Now For Authors Call for Submissions Statistics Contact
Home > Archives > Volume 20, No 11 (2022) > Article

DOI: 10.14704/nq.2022.20.11.NQ66262

An Effectual Hybrid Feature Selection Method Towards the Classification Performance with Support Vector Machine

Arti Patle, Gend Lal Prajapati


Digitization has revolution in this era; the way person communicates and receives information. Intelligence and prediction combined for enlighten the solution for e-commerce, medical, agriculture etc. Machine learning gives the solution for prediction and inference knowledge, Support Vector Machine (SVM) is the best Classification model for that. Recently, databases have rapidly grown their dimension in all domains, high dimension of input dataset vector also degrade the performance of classifier. So, the feature selection (FS) is the best solution for high dimensional dataset, it is the important work for discover the knowledge with data mining, data analysis, machine learning and pattern analysis. Since, it is very helpful to prediction quality improvement, computational time reduction and model simplification. In this research work we propose formation of a new Hybrid Filter- Wrapper feature selection method for selection of appropriate input feature. Real world problem with anonymous search space handle by this method very efficiently. Proposed method is Hybrid Filter-Wrapper method gives good subset of feature, so the kernel parameter already optimized with it. The proposed method jointly apply ranking and best first search. We applied the proposed method on different domain dataset for classification. Experimental result of this paper shows that proposed approach correctly select and identify the efficient, relevant feature and achieve best classification accuracy.


Support Vector, Classification, Inference knowledge, Feature Selection

Full Text